Methodology and data sources
PDI publishes its models rather than asserting authority. Every figure in a report is reproducible from the structured record shown on that report, using the formulas below. Model version pin-1.1.0.
Valuation model
Automated valuation and range
- · Comparable sales within 1.2 miles over a trailing 180-day window are adjusted for living area, bathroom count and construction vintage.
- · The adjusted comparable $/sqft set is blended with the submarket base $/sqft to produce a point estimate, then scaled by living area.
- · The published range is a function of comparable dispersion: wider dispersion widens the interval and lowers stated confidence. Confidence is a dispersion measure, not a probability of sale price.
- · Output is a modeled estimate. It is not an appraisal, broker price opinion or a certified valuation, and it should not be relied on for lending or transactions.
Rent, yield and operating cost
Income modeling
- · Market rent = submarket rent per square foot × living area, bounded by a band.
- · Monthly operating cost includes effective property tax, insurance, HOA dues and a vacancy provision. Maintenance, capex reserves, management fees and financing costs are excluded in V1 and must be added by the user.
- · Gross yield = annual gross rent ÷ valuation estimate.
- · Net yield = (annual gross rent − annual operating cost) ÷ valuation estimate.
- · Cap rate uses net operating income on the same modeled cost basis.
PDI Score
Five weighted components · pin-1.1.0
| Component | Weight |
|---|---|
| value stability | 22% |
| rental yield | 23% |
| liquidity | 20% |
| supply pressure | 15% |
| risk exposure | 20% |
Components are normalized to 0–100 and combined by weight. The score ranks relative modeled attractiveness inside PDI's dataset. It is not a credit score, an appraisal, or a buy/sell recommendation, and it is versioned so historical scores remain interpretable.
Forecast scenarios
12 / 24 / 36 months
- · Forward paths compound modeled annual submarket growth, damped by months of supply and days-on-market percentile.
- · Interval width scales with the square root of the horizon to reflect compounding uncertainty.
- · Forecasts are scenarios, not predictions or guarantees. They exclude interest-rate shocks, policy changes, natural disasters and idiosyncratic asset events.
Data sources and provider architecture
Transparency
- · This build runs on a labeled synthetic demo dataset. Demo output is never presented as live, and every report carries a persistent data-mode badge.
- · Provider access is abstracted behind a server-side registry with read-through caching. Live providers appear as stubs until credentials are configured; no secrets are required for the demo to function.
- · Historical snapshots and score versions are persisted server-side so past reports remain auditable against the model that produced them.
- · No provider keys or secrets are ever exposed to the browser.
Important disclosures
Property Data Intelligence produces modeled estimates, not appraisals, broker price opinions, or investment advice. Valuations, rent estimates, scores and forecasts are generated by deterministic statistical models from the data described on each report. Forecasts are scenarios that assume published trend decay; they are not guarantees of future value or return.
In this build all figures come from a labeled synthetic demo dataset. Demo output is never presented as live data. PDI is not a lender, broker, appraiser, or consumer reporting agency, and outputs may not be used for eligibility decisions covered by the Fair Credit Reporting Act or for Fair Housing–regulated screening.
Verify all figures independently before transacting. Read the methodology and data-source detail.